AI will surpass peak human intelligence across ALL economically valuable domains by 2030, extending the sequential maturation roadmap beyond 'independent researcher' (2028) into full-domain superintelligence.

AI will surpass peak human intelligence across ALL economically valuable domains by 2030, extending the sequential maturation roadmap beyond 'independent researcher' (2028) into full-domain superintelligence. | Frontier model expert-level GDPval saturation

Key catalyst: Frontier model expert-level GDPval saturation

Watch events: Frontier model expert-parity benchmarks; GDPval saturation; OpenAI capability disclosures

Altman sequential roadmap internally consistent 2023-2026; GPT-6-class capabilities forecast for 2027-2028. Peak-human ASI by 2030 aligns with Aschenbrenner intelligence-explosion framing but aggressive vs Hassabis / Karpathy.

{
  "nia": false,
  "qty": "peak-human ASI",
  "mode": "FORECAST",
  "role": "Cited-CEO",
  "context": "Extends Altman's prior intern-by-2026 (CMQ_001) / researcher-by-2028 (CMQ_002) sequential-maturation framing into ASI tier. Altman has also publicly distanced OpenAI from the term 'AGI' as 'not a super useful term' due to definitional ambiguity.",
  "to_year": 2030,
  "conv_cues": "CEO FIRST_PERSON; explicit year; deterministic scaling-law framing",
  "direction": "HAPPEN",
  "from_year": 2030,
  "timeframe": "by 2030",
  "conv_level": "HIGH",
  "milestones": [
    {
      "kind": "llm_pre_event",
      "label": "Frontier model crosses 80% on full GDPval gold set across all 9 sectors",
      "source": "https://artificialanalysis.ai/evaluations/gdpval-aa",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -9,
      "source_id": null,
      "confidence": 0.75,
      "source_url": "https://openai.com/index/gdpval/",
      "expected_date": "2027-01-30",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2027-06-30",
        "from": "2026-09-01"
      },
      "measurement_criterion": "GPT-5.4 already 83% per Artificial Analysis; signal is sustained >80% expert-tie-or-win rate across nine sectors with public reproducibility"
    },
    {
      "kind": "llm_pre_event",
      "label": "Public AI agent autonomously runs an end-to-end biology, chemistry, or ML research project (no human checkpoint > 24h)",
      "source": "Internal estimate based on Anthropic Claude agent / OpenAI Operator / Sakana AI Scientist trajectories",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -8,
      "source_id": null,
      "confidence": 0.45,
      "expected_date": "2028-01-01",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2028-12-31",
        "from": "2027-01-01"
      },
      "measurement_criterion": "Verifiable demonstration: AI agent runs >=14-day continuous research project producing publishable result, with logged human intervention < 24h granularity"
    },
    {
      "kind": "llm_pre_event",
      "label": "Frontier-lab AI safety institute reports model exhibiting 'beyond peak human' performance in 3+ economically significant domains",
      "source": "Internal estimate based on AISI evaluation cadence",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -7,
      "source_id": null,
      "confidence": 0.5,
      "expected_date": "2028-06-15",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2029-06-30",
        "from": "2027-06-01"
      },
      "measurement_criterion": "US AISI, UK AISI, or METR formal evaluation report identifies superhuman performance (top decile of human experts) in >=3 of {software engineering, mathematical research, drug discovery, financial analysis, legal analysis}"
    },
    {
      "kind": "scenario_signal",
      "label": "Scenario fires: AGI mid: Kurzweil 2029 path",
      "status": "pending",
      "weight": 0.7,
      "ordinal": -6,
      "source_id": "S_AGI_MID_2029",
      "expected_date": "2029-03-31",
      "observed_date": null
    },
    {
      "kind": "llm_pre_event",
      "label": "Industry expert blind evals show AI win rate >50% across all 44 GDPval occupations",
      "source": "https://openai.com/index/gdpval/",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -5,
      "source_id": null,
      "confidence": 0.4,
      "expected_date": "2029-06-15",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2030-06-30",
        "from": "2028-06-01"
      },
      "measurement_criterion": "GDPval (or successor) public leaderboard: leading model achieves >50% strict 'win' (not tie) in blind expert review across all 44 occupations"
    },
    {
      "kind": "llm_post_event",
      "label": "Cascade: Sam Altman publicly declares OpenAI internal model is 'peak human' across all economically valuable domains",

... (truncated)